A 0.5B parameter coding model fine-tuned to think before it codes — specialising in bug
analysis, code review, algorithm problem-solving, and agentic planning.
Built by Sandeep Reddy · TIMPS · Made in India 🇮🇳
ollama pull sandeeprdy1729/timps-coder
ollama run sandeeprdy1729/timps-coder
Python (Transformers)
fromtransformersimportAutoModelForCausalLM,AutoTokenizermodel=AutoModelForCausalLM.from_pretrained("sandeeprdy1729/TIMPS-Coder-0.5B")tokenizer=AutoTokenizer.from_pretrained("sandeeprdy1729/TIMPS-Coder-0.5B")messages=[{"role":"system","content":"You are TIMPS-Coder v3. THINK through the root cause, FIX with complete code, VERIFY edge cases."},{"role":"user","content":"Fix: `data['user']['email']` throws KeyError when email is absent."},]text=tokenizer.apply_chat_template(messages,tokenize=False,add_generation_prompt=True)inputs=tokenizer(text,return_tensors="pt")out=model.generate(**inputs,max_new_tokens=700,temperature=0.1,do_sample=True)print(tokenizer.decode(out[0],skip_special_tokens=True))
MLX (Mac Apple Silicon)
pip install mlx-lm
mlx_lm.generate \
--model sandeeprdy1729/TIMPS-Coder-0.5B \
--max-tokens 700 --temp 0.1 \
--prompt '<|im_start|>system
You are TIMPS-Coder v3. THINK through the root cause, FIX with complete code, VERIFY edge cases.<|im_end|>
<|im_start|>user
Fix the race condition: two threads increment self.count += 1 simultaneously.<|im_end|>
<|im_start|>assistant
'
Training Details
Fine-tuning Configuration
Parameter
Value
Base model
Qwen/Qwen2.5-Coder-0.5B-Instruct
Fine-tuning method
LoRA (Supervised Fine-Tuning)
LoRA rank
16
Learning rate
5e-6
Iterations
3,000
Batch size
1 (grad accum ×4)
Max sequence length
2048 tokens
Framework
MLX-LM on Apple Silicon
Peak RAM
~5.5 GB
Training Data
Dataset
Type
Approx. Samples
newfacade/LeetCodeDataset
Algorithm problems with solutions
~2,500
SWE-bench/SWE-bench_Verified
Real GitHub issue → patch
~400
TIGER-Lab/SWE-Next-SFT-Trajectories
Agentic edit traces
~2,000
WaltonFuture/agentic-sft-new
Tool use + bash planning
~3,000
Custom TIMPS bug-fix corpus
Hand-curated bug/fix pairs
~500
Total
~8,400 samples
All samples formatted in ChatML with THINK → FIX → VERIFY answer structure.
-brightgreen){kind=icon}